Posted by Chris on May 02, 1998 at 05:38:27:
In Reply to: Found it? Please help, rpcman, Chris posted by Walker on May 01, 1998 at 14:38:08:
: I found frequent mention of this species "Drosophila melanogaster"
: in the list of titles. What is it and why is it studied so often?
Drosophila is a fruit fly. It is a common genetic model because it is fairly complex, yet it is easy to handle in the laboratory and has a rapid generation time. Sudying humans or redwoods would be much less convenient.
: Here are a few that sounded like possibilities of what I am looking for. Please
: comment if you are familiar with these or know of any other study
: which deals with the predictability of evolution. (Not simply that evolution
: will happen -- not very useful) Predictions of how, when and why
: specific examples of evolution (not gene shift) will happen.
A few posts back, you suggested taking a simple organism with a rapid generation time, placing it in a controlled environment, and tracking its evolution under changing stress. There are a couple of classical papers that used this kind of experiment to determine, among other things, how the structural proteins of a bacteriophage interact. Incidentally, I think they illustrate some of the predictability you are asking for:
Jarvik, J. & Botstein, D. (1975) Conditional-lethal mutations that suppress genetic defects in morphogenesis by altering structural proteins. Proceedings of the National Academy of Sciences, USA 72:2738-2742
Floor, E. (1970) Interaction of morphogenic genes of bacteriophage T4. Journal of Molecular Biology. 47:293-306
Both of these studies used bacteriophages--viruses that infect bacteria--which are so simple that all of their genes were known even in the 1970s. Some of their genes code for proteins necessary for biochemical survival and reproduction within the host bacteria, while the remainder code for proteins that make up the structure of the completed virus (morphogenic genes). Both studies used strains of virus that were defective in making their structural proteins and that could only reproduce within certain permissive strains of E. coli. By taking a population of mutant viruses and forcing them to infect non-permissive bacteria, so-called revertant mutations could be produced with some predictability.
A small percentage of the new mutations simply corrected the mutation in the virus and returned it to the wild-type strain. More often, the suppressor mutation occured in a gene for another structural protein that ineracted with the mutated one (imagine buying the wrong sized screws--you can either change the screws to the right size, or you can change the holes to fit the screws). Mutating a biochemical gene, such as one that replicates the viral DNA, was of no benefit to the virus with defective structural proteins: only mutations occurring in proteins associated with the originally mutated structural protein would allow survival. Often, the new mutant viruses were now sensitive to cold or to heat. By taking a new cold-sensitive mutant and forcing it to grow in the cold, a second generation of mutants could be derived that were now cold permissive but heat-sensitive instead. Successive original mutations produced a flip-flop between cold sensitive and heat sensitive generations.
From these studies we can see that though mutations occur at random, there is a certain predictability to the kind of mutations that give a selective advantage.
: Chris says, "Science doesn't do proof--that's mathematics. Science only deals with evidence and
: experimentation to reach its conclusions. You are asking for more than science can offer--even
: Newton's Laws (which definitely aren't universal, by the way)."
: I always liked Newton's laws because they are so powerful in terms of
: predicting what will happen in the future. Every day we can make predictions
: that come true. Because of them (modify them for Einstien if you like)
: the sun rises, Planets move, if I throw a rock, it will follow a path
: predetermined by gravitational laws, every time. You simply can't deny it.
: Is anyone uncomfortable with stating a distinction between evolution and
: other sciences in terms of predictability?
Actually, I would feel better if you didn't single out evolution. Newtonian physics deals with pretty simple systems under idealized conditions. One object moving under known forces strikes another object moving under other known forces and by plugging the known data into your formula you can make a very precise prediction. But can you do that with most fields of science, or even with all physical problems? What if you have ten moving objects striking each other? Meteorology basically boils down to physics. It deals with temperature, movement of air, humidity, etc. all of which are forces that can be modeled by physics, yet nobody can predict the weather outside of a few rough, short-term generalities because the system is way too complex. Geologists have identified the regions where contintental masses intersect and grind against one another--why can't they predict when the next earthquake or volcanic erruption will occur? Shall we haul him in for pseudoscience because the best he can do is say: "an earthquake is likely to happen here." Your doctor, knowing all of your genetic background and lifestyle risks can only inform you that you are more/less likely than average to develop heart disease or colon cancer--he can't give you an idealized prediction. The evolution of a complex organism deals with a dash of random mutation over a hugely complex system of interacting genes and environmental factors. Given the complexity of the problem, what level of prediction should you expect?
Prediction is a great goal for good science, but I wouldn't use a certain level of prediction as a determinant of what is or is not a science. Maybe we could say that it is one way to measure how refined is an area of science.
: It certainly doesn't mean that
: evolution is less true, just less provable. Nature simply doesn't afford
: the luxury of predictability to evolution that it does to other
: sciences.
Or, nature doesn't afford to other sciences the level of predictibilty that Newtonian physics has attained with certain ideal systems.